Model Description
This model is designed to eliminate the "oily/glossy skin" texture often found in Illustrious checkpoints, delivering a Furry Bara checkpoint rich in artistic flair. You can expect images with exceptional lighting and distinct stylistic qualities.
Merge Process / Recipe
Base Construction: Performed a layered merge of Indigo_Furry_Mix_06 and ShadowForge to create the core
Model 1.Style Extraction: Integrated style LoRAs into Red Lily, then subtracted Illustrious 0.1 to extract a pure
Style-Oriented LoRA.Final Assembly: The
Style-Oriented LoRAwas layer-pruned and merged intoModel 1alongside Kombagavad_Gita3.9 and NAI_vpred_fix at specific ratios.
Update Notes
⚠️ Important Note: Each version in this repository is designed with a distinct focus. A higher version number does not necessarily mean it is "better" or the right fit for your specific needs. Please choose the version that best aligns with your creative goals.The appropriate negative tags vary from version to version.Please do not over-tweak these negative prompts, and avoid blindly pasting "legacy" negative prompt blocks that you are used to using on other models. Using incorrect or bloated negative prompts on this version may actually break the model's ability to properly fix anatomical structures and composition errors.
Version 4.0
In this version, I have significantly adjusted the merging formula and shifted the foundational base model to cknoob (Chenkin Noob XL ckxl).
The primary objective of this update is to drastically weaken the heavy baked-in base style of previous versions. This makes the model much more versatile, flexible, and responsive to your specific prompts and LoRAs. However, as an unavoidable trade-off of this generalization, you may occasionally encounter the smoother, somewhat "oily" or glossy skin textures commonly seen in other standard models.
Merge Process / Recipe:
Style Extraction: Extracted a difference LoRA from ShadowForge by pruning it against Noob 1.1.
Base Injection: Merged this newly extracted ShadowForge LoRA into cknoob 0.2.
Final Assembly: Performed a precise layered block-merge between the resulting model and Indigo 11.
Version 1.5 (Generalization & Compatibility Focus)
Built upon the foundation of v1.0, this version intentionally dials back the heavy stylization to offer a much more versatile and generalized base. The generation experience feels closer to traditional foundational models, providing significantly improved compatibility with other community resources (like your favorite style LoRAs). Additionally, v1.5 excels at handling complex compositions, performing well with negative space, wide-angle perspectives, and panoramic shots.
Version 3.0 (Concept & Character Focus)
This version doubles down on Furry-related concepts, pushing the boundaries of character anatomy while actively restructuring the underlying stylistic composition. Because of its intense focus on the subject, v3.0 is recommended for artworks where the character is the absolute focal point and occupies a large portion of the frame (such as portraits, close-ups, or dynamic character-driven scenes).
Version 2.1
Anatomy Optimization: Significantly improved anatomical correctness for more stable and natural posing.
Enhanced Furry Concept: Strengthened the model's understanding of furry aesthetics and specific traits.
Higher Success Rate: Increased the "good image" ratio, making it easier to get high-quality results with fewer prompts.
Style Shift: Please note that these improvements have led to a slight evolution in the base artistic style compared to v1.0.
Regarding Version 1.0 (Refined) If you still prefer the aesthetic of the original version, I have re-uploaded a fixed v1.0 (TE Fix).
TE Fix: I repaired the Text Encoder, which was likely the cause of previous high error rates and poor LoRA compatibility.
The core functionality remains the same, but it should now be much more stable.
⚠️ 1.0 Version:
Due to the heavy stylization inherent to this model, you may occasionally encounter anatomical irregularities, such as missing limbs. The model is also quite sensitive to prompting. For the best user experience and significantly improved stability, I highly recommend pairing this checkpoint with your favorite Style LoRAs.
Recommended Settings
Quality Tags:
4.0:
Positive Prompt: masterpiece, best quality, very awa, newest, highly detailed, excellent composition, fully rendered, expressive, intricate details
Negative Prompt: worst quality, worst aesthetic, jpeg artifacts, bad anatomy, bad proportions, fusion, conjoined, intersecting, distorted, deformed, mutated, malformed hands, poorly drawn hands, bad hands, disconnected limbs, extra limb, malformed limbs, watermark, signature, text, disfigured, fused eyes, female, girly, noisy texture, cluttered details, distortion, flexible deformity, fused fingers, ai-generated, bad feet, bad tail
Positive(1.0-3.0):
masterpiece, very aesthetic, best quality, very awa, amazing quality, absurdres, newest, high detailNegative(Please select the appropriate version of tags to use):
1.0:
worst quality, worst detail, worst aesthetic, flat color, jpeg artifacts, distortion, ai-generated, scan artifacts, malformed hands, cluttered details, noisy texture, watermark, signature, text, female, girly, bad anatomy, bad proportions, ugly, fusion, mutated, mutation, deformed, distorted, (missing limb:1.2), (disconnected limbs:1.1), (malformed limbs:1.1), flexible deformity, disfigured, fused eyes, skinny, broken tail, bad tail2.1:
worst quality,worst detail,worst aesthetic,flat color,jpeg artifacts,distortion,ai-generated,scan artifacts,malformed hands,cluttered details,noisy texture,watermark,signature,text,female,girly,bad anatomy,bad proportions,(unclear pupils and eyelids:1.1),ugly,error,fusion,mutated,deformed,distorted,bad hands,fused eyes,flexible deformity,malformed limbs,disconnected limbs,disfigured,skinny,bad feet,bad tail1.5:
worst quality,worst aesthetic,jpeg artifacts,signature,watermark,text,bad anatomy,bad proportions,deformed,disfigured,mutated,(missing limb:1.1),distorted,disconnected limbs,malformed limbs,bad hands,(mutated hands:1.1),fused eyes,fusion,conjoined,intersecting,female,girly,skinny,lean,slim,lanky,boyish,slender,flat color,noisy texture,cluttered details,distortion,flexible deformity,extra limb,ai-generated,bad feet,bad tail,low quality3.0:
worst quality,worst aesthetic,jpeg artifacts,signature,watermark,text,bad anatomy,bad proportions,fusion,deformed,disfigured,mutated,distorted,disconnected limbs,malformed limbs,missing limb,bad hands,mutated hands,fused eyes,conjoined,intersecting,female,girly,skinny,flat color,noisy texture,cluttered details,distortion,flexible deformity,ai-generated,bad feet,bad tail,low quality,
Lighting & Style Tags:
chiaroscuro, Natural Light, Mood Lighting, ray tracing, dappled lightDefault Style Enhancers:
cinematic lighting, volumetric lighting, dramatic shadows, impressionism, blurred background, atmospheric, depth of field, dynamic angle
Generation Parameters:
CFG Scale: 3.5 - 7
Sampler: Euler a + SGM Uniform OR DPM++ 2M + Karras
Steps: 22+
Upscaler: 001_classicalSR_DF2K_s64w8_SwinIR-M_x4 OR 4x_foolhardy_Remacri
Useful Resources:
(https://civitai.com/articles/8817/illustrious-year-modifier-tags)
(https://drive.google.com/file/d/1FuME-Ch5a9PsfDX5DN68ygjQaR28EdXZ/view?usp=drive_link)
Credits & Special Thanks
This model would not be possible without the incredible work of the community. I would like to express my sincere gratitude to the creators of the following models and LoRAs for their innovation and high-quality resources used in this merge:
Indigo Furry Mix XL – View Model
ShadowForge-NAI_XL – View Model
Red Lily | Illu – View Model
Chromatic Trip – View Model
Umehara Sei Style (Illustrious XL) – View Model
Modern Chinese Painting Style (Illustrious) – View Model
NTDMIX – View Model
NAI Vpred Fix / Utility – View Model
Kombagavad Gita - Furry Bara Style – View Model
Illustrious XL – View Model
Chenkin Noob XL (CKXL) – View Model
Many thanks to Soxan for his guidance and support. Please support this talented creator: @Soxan
感谢来自包括白轩在内的诸位友人的提示词援助。
Terms of Use for Model
Thank you for your interest in using this model! To ensure a positive and fair experience for everyone in the community, I’ve outlined the following terms. These are designed to protect both you as a user and me as the creator, while encouraging creative exploration. Please take a moment to read through them before using the model.
1. Usage Restrictions
This model is provided for creative and personal use. However, it must not be used for purposes that violate laws, ethical standards, or community guidelines. Specifically, the model is prohibited from being used for:
Creating or distributing content that is illegal, harmful, or offensive. Used to create any infringing content.
Any form of exploitation, harassment, or discrimination.
I do not authorize any third-party claims of copyright on the style generated by this model.
Commercial purposes without prior written permission from me.
Note: These restrictions are in place to maintain a safe and respectful environment for all users.
2. Copyright and Licensing
I retain full copyright ownership of this model. You are granted a non-exclusive, non-commercial license to use the model for personal or artistic purposes. Commercial use or redistribution of the model itself is not permitted without my explicit consent.
There are no restrictions on open-source merged models being used for image generation services, but it is recommended to credit the sources of any merged models.
I will not be responsible for any consequences resulting from the unauthorized reproduction of this model on platforms other than Civitai.
3. Limitation of Liability
I am not responsible for any damages, losses, or issues arising from the use of this model. This includes, but is not limited to, unintended outputs, data loss, or compatibility issues with your hardware or software. I encourage you to exercise your own judgment when using the model and to ensure it meets your specific needs.
4. Use at Your Own Risk
By using this model, you acknowledge that you do so at your own discretion and risk. I cannot guarantee the model’s performance, outputs, or suitability for any particular purpose. Please ensure you understand the potential risks and use the model responsibly.
By using this model, you agree to abide by these terms. If you have any questions or need clarification, feel free to reach out to me through Civitai’s messaging system. I’m excited to see what you create, and I hope this model inspires your work!